You've successfully subscribed to Better Data Science
Great! Next, complete checkout for full access to Better Data Science
Welcome back! You've successfully signed in
Success! Your account is fully activated, you now have access to all content.

3 Reasons Why You Shouldn't Become a Data Scientist

3 Reasons Why You Shouldn't Become a Data Scientist

Data Science is fun, but is it a career you should choose? Think twice if any of these 3 reasons hit home.

Data Science is somewhat of a misunderstood field. On paper, you're solving global warming and cancer straight of out college for a mid 6 figure salary. In reality, you're struggling to understand why your SQL joins are duplicating the data. Both examples are completely made up, but they serve perfectly to illustrate a point - it's essential to manage expectations.

What can you expect from a data science job? Can you get hired as a junior in a small company? Are salaries really that high or are the numbers skewed by a huge number of seniors in the field? These are all great questions I won't answer today.

Instead, let's focus on some practical reasons why data science might not be for you.


Reason #1: You Want Your Work to be Finished

How do you know your job is done as a data scientist? You don't.

Consider data engineers. In oversimplified terms, they build pipelines that move data from one place to the other. Once the data is moved, their job is done.

Web developers are another good example. If a full-stack engineer needs to add a button to the app, he/she knows exactly what has to be done to, well, add the button and connect it to a set of functionalities. Once the button is implemented and tested, their job is done.

But what about data scientists? It's not so easy to mark your job as done in this field. The predictions might work fine today, but tomorrow something unexpected happens and you're back to square one. It's a more abstract field because you typically never know how someone will use your model. You can't think of every possible scenario beforehand.

Also, your model is only as good as the set of data you have at the moment. If you collect more data over time, it usually means you can improve on what you currently have in production.

Data science projects are hence more cyclical than linear. It's not necessarily a bad thing, but some people prefer to have their tasks completed at the end of the day.

Reason #2: You Want to Build Something Tangible

Do you know how a mobile app developer can explain what they do for a living? They just grab their phone and show the app, or at least redirect people to the app store. The same goes for web developers and other IT professionals that build something tangible for the end user.

Data science feels more like a small cog in a big machine. I spent months working on a project only to have the model results displayed in a single box on a huge dashboard. Sure, you could drill down into the model results, but that's a tangible feature someone else built on top of my work.

As a data scientist, you're far more likely to serve as a connecting point between raw data and presentable insights. Someone else supplies the raw data, and someone else makes presentable insights. You're essentially a middleman between data engineers and software engineers.

Valuable? Certainly. Tangible? Not so much.

Reason #3: You Don't Want to Keep Up with the Trend and Research

Any engineering job implies constant learning as technologies evolve, no arguing there. But data science takes that to a whole new level.

Researchers in data science and machine learning are constantly coming up with new algorithms or improvements to previous iterations. Take YOLO (You Only Look Once) object detection algorithm for example. The original implementation dates back to 2015. In late 2022, we have YOLO, PP-YOLO, YOLOP, YOLOX, YOLOF, YOLOS, and YOLOR, each having their own versions. For example, the plain YOLO algorithm versions range from V1 to V7, each bringing something new to the table.

It's just a single family of algorithms designed to tackle a very specific computer vision problem. Don't even get me started on BERT.

Of course, you don't need to know everything, but the field can feel overwhelming at times due to a ton of research that's been going on in the last decade.


So, Is Data Science for You?

Data Science and Machine Learning aren't going anywhere. But are these fields right for you? It's a question only you know the answer to.

It's important to consider both pros and cons. There are numerous articles/videos listing the pros, so I wanted to make an entire article showcasing the not-so-sexy aspects of the job. Maybe you're okay with all three reasons I listed - I know I am - but they can be huge energy drainers and nd motivation killers later down the road.

There's no harm in trying, you can always go back.

Have you transitioned from software engineering to data science? What are the things you miss about your old job? Let me know in the comment section below.

Stay connected